INDEX

Explanations

words related to doors and interactions with doors

oai_token-act-pair · gpt-3.5-turbo

references to doors and their actions or symbolism

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

conom

-0.94

luster

-0.81

 Saga

-0.80

 Samoa

-0.78

 endorsements

-0.77

nutrition

-0.69

Buff

-0.67

aukee

-0.67

 Parenthood

-0.67

rities

-0.66

POSITIVE LOGITS

 doors

1.74

 door

1.54

 entrances

1.52

 Doors

1.42

 doorway

1.39

 stair

1.39

 Door

1.36

 entrance

1.35

doors

1.33

door

1.32

Activations Density 0.409%

words related to doors and interactions with doors

references to doors and their actions or symbolism

No Comments

No Known Activations